Page MenuHomePhabricator

Improvements to Annotool
Closed, ResolvedPublic

Description

In this task we list improvements and bugs to be solve on the Annotool.
Please report bugs here.

  • Remove list of wiki_db (top right)
  • Remove edits to sandbox
  • T343973: Fix relative links for Qids in Annotool
  • Allow to edit annotations
  • Revise how Qid label and description changes are indicated in relation to the review of statement edits (see comment below with example)
  • Add a license in the repo

Event Timeline

In the description you list "Remove list of languages". Instead of removing the drop-down box, can it not instead be made functional, i.e. only display tasks in the chosen language? For example, I'm only proficient in English, so it is likely better for me to focus on tasks in English, rather than additionally displaying Spanish, German, etc. Being able to choose which language these training tasks are from would be useful.

In the description you list "Remove list of languages". Instead of removing the drop-down box, can it not instead be made functional, i.e. only display tasks in the chosen language? For example, I'm only proficient in English, so it is likely better for me to focus on tasks in English, rather than additionally displaying Spanish, German, etc. Being able to choose which language these training tasks are from would be useful.

That filter is not really language but wiki_db, and it's designed for evaluating revision on Wikipedia projects. I'll update the description, thanks.

It'd be great if there were an "undo" button after you evaluate a given edit, in case you misclick.

A revision of the review scope or indication of changes may be needed to better address what may lead to false annotations for hijacked Qids where the label and description have changed prior to the statement that is being compared. At face value the statement change looks like it is a simple mark as a revert, but viewing the edit history there was a series of edits which changed many statements changing the entire context of the Qid.

For instance: https://www.wikidata.org/wiki/Q117409403?oldid=1918143489

Screenshot from 2023-08-14 20-04-52.png (488×2 px, 36 KB)

and

Screenshot from 2023-08-14 20-04-30.png (514×2 px, 37 KB)
https://www.wikidata.org/wiki/Q117409328?oldid=1922750462 .

Hi @wolfgang8741

Thanks for the input. In order to make the annotations as much useful as possible for training the algorithm, it would be good to have labels that are not too specific, and can generalize as much as possible. Do you think that something like: "long-term vandalism" or "hijacked item" could be a good name for the phenomena that you are describing ?

Hi @diego

I think "hijacked item" makes more sense as the label, long-term vandalism to me sounds more like the item is frequently vandalized rather than an instance of hijacking (which could occur multiple times in the history of an item).

My previous comment was aimed at raising that there may be a need for thinking about what types of cues and information are presented to the users via the Annotool UI that would aid in detecting such a hijack and encourage investigating the Qid history instead of relying on the Name and Label presented. It may be just as simple as noting the label and/or description was changed by the same user for which the statement under review is being checked, or adding a count of the number of times the label and/or description changed since some point in time, or how long it has been since the labels field and/or how long since the description was changed before the statement being assessed was edited. A more recent edit may provide an indication of the inertia of the Qid is lower and may need for more review than if the label had been there for months or never changed.

It may be worth adding to the existing links a "Qid history" button ie to https://www.wikidata.org/w/index.php?title=Q117409403&action=history on the same lines as the label following the existing buttons like (1901202637) button in this image. as (Qid history)

Screenshot from 2023-08-15 21-51-26.png (90×636 px, 10 KB)
followed by a [Label last edited: less than 24 hour prior], [Description last edited: 3 days prior] and maybe the same last edited for the description. If there is concern of overloading the UI, the "last edit" could just be shown on hover on the label and description as it probably is not needed for most reviews.

KHernandez-WMF set Due Date to Dec 29 2023, 6:00 AM.

@MunizaA , could we please add an action to finish a project? By finish I mean to keep the project data, but stop showing in the front-end.

Aklapper renamed this task from Improvements to Annotool to Improvements to Annotool.Oct 26 2023, 8:43 AM

@MunizaA / @KHernandez-WMF : Hi, the Due Date set for this open task passed a while ago.
Could you please either update or reset the Due Date (by clicking Edit Task), or set the status of this task to resolved in case this task is done? Thanks!

Aklapper removed Due Date which was set to Dec 29 2023, 6:00 AM.Jul 22 2024, 6:59 AM

@diego the work needed to resolve this task is trivial but as far as I know, we're not running any campaigns on Annotool at the moment. If that is indeed the case, would it make sense to move this to the freezer and pick it back up once that changes? Thanks!

Moving to freezer per Muniza's comment above. If that changes, please bring it back to Research-->Backlog lane for prioritization.

leila edited projects, added Research-Freezer; removed Research.
leila added a subscriber: MunizaA.
diego updated the task description. (Show Details)